Looking for more info on 'max-distance' argument of KYCG_buildGeneDBs function

[bethan-mallabar-rimmer]

The KYCG_buildGeneDBs function has a default argument of max_distance = 10000, defined in the documentation as "probe-gene distance for association". I interpreted this as any site 10,000bp up or downstream of a gene will be considered associated with the gene. However, I'm not sure if this is correct (and it also seems like quite a long distance for association).

Reducing max_distance from 10000 to 200, built less gene databases (232 rather than 1757) but found more significant genes (229 rather than 99 with p < 0.05 and FDR < 0.05), presumably because p value correction is less stringent when looking in less databases.

Can anyone provide more info on what max_distance means please or recommended input?